Did my recent ancestor who failed high school classes have math talent or not? And did my ancestors who didn't have a high school at all but could build a barn that still stands one hundred years later have math talent or not? How can we even talk about whether this thing is inheritable if we don't define it well and it manifests in such different ways in different environments?
The simplest assumption is that these kids are bad at math because math in our elementary and middle schools sucks! Look at the stats for immigrants to the US. If you came from impoverished places without educational opportunities, educational attainment rises over time. If you are a kid who comes from Hungary or Romania or Russia or South Korea, you'll be so much better at math than the US kids that it's not even funny. If your grandparents came from Hungary or Romania or South Korea and kicked ass at math, but your family has been in the US for a few generations, you'll suck as much as any other average American. If you're going to push heritability, you're going to have to face the conclusion that the water and air in America prompt genetic mutations that make people bad at math over a few generations. (I'm so glad I'm first-gen in part, but I fear for potential children.)
"The more the math-anxious parents tried to work with their children, the worse their children did in math, slipping more than a third of a grade level behind their peers."
The implication is that if the math-anxious parents did not work with their children as much, these children did better.
That's useful "know your limitations" information, but I don't see how it separates nature vs nurture effects.
Math talent, like other intelligence, is inheritable. So until you've pried apart nature and nurture, the simplest assumption is that these kids are bad at math because they don't have much talent for it.
Google Turkheimer's Three Laws of Behavior Genetics for more on this often neglected field.